Blue Sky Thinking UoW-WWT Final Report_FINAL_Web

Blue Sky

Thinking: Evaluating the Impacts of Volunteering in Wetland Nature on Creative Thinking and Innovation

A report supported by Bank of America Paul T. Sowden, Alexander Smith, Jean-Christophe Goulet-Pelletier, Jonathan Reeves, Georgie Cox, Phoebe Fellows, Isabelle Kauer, George Murrell, Emily MacDonald, Lorna Jarvis & Birgitta Gatersleben

Blue-Sky Thinking: Evaluating the Impacts of Volunteering in Wetland Nature on Creative Thinking and Innovation A report supported by Bank of America Paul T. Sowden1, Alexander Smith2, Jean-Christophe Goulet-Pelletier1, Jonathan Reeves3, Georgie Cox3, Phoebe Fellows3, Isabelle Kauer1, George Murrell1, Emily MacDonald1, Lorna Jarvis1 & Birgitta Gatersleben4 1. School of Psychology and Social Sciences, University of Winchester, UK 2. School of Psychology, University of Plymouth, UK 3. Wildfowl and Wetlands Trust 4. School of Psychology, University of Surrey, UK Suggested citation: Sowden, P. T., Smith, A., Goulet-Pelletier, J., Reeves, J., Cox, G., Fellows, P., Kauer, I., Murrell, G., MacDonald, E., Jarvis, L. & Gatersleben, B. (2026). Blue-Sky Thinking: Evaluating the Impacts of Volunteering in Wetland Nature on Creative Thinking and Innovation. University of Winchester, UK.

Contact information Lead researcher: Prof. Paul Sowden, paul.sowden@winchester.ac.uk WWT project lead: Dr Jonathan Reeves, jonathan.reeves@wwt.org.uk

Acknowledgements This research was supported by funding from Bank of America together with support from Employee Volunteering to access non-nature volunteer participants.

Contents List of Figures and Tables

1

Executive Summary

3

1.0

Introduction

6

1.1

Creativity and Innovation

6

1.2

Nature Creativity Connection

7

1.3

Present Research Aims and Hypotheses

7

2.0

Method

8

2.1

Participants

8

2.2

Design

8

2.3

Measures

8

2.3.1

Primary Outcome Measures

8

2.3.2 Secondary Outcome Measures

9

2.4

Procedure

9

3.0

Results

12

3.1

Plan of Analyses

12

3.2

Descriptives

12

3.3

Manipulation Check

12

3.4

ANCOVA Assumptions

13

3.5

Primary Outcomes Results

14

3.5.1

Creative Self-Efficacy

14

3.5.2 Top 2

14

3.5.3 Fluency

15

3.5.4 Connection Test Results

16

3.6

Secondary Outcomes Results

16

3.6.1

Nature Integration

16

3.6.2 3.6.3 3.6.4 3.6.5 3.6.6 3.6.7 3.6.8 3.6.9 3.7

3.8

Arousal Valence Control State Openness State Inspiration Perceived Restorativeness Inspired by the Environment Inspired by the Activity Exploratory Regression Model: effects of inspiration and restoration on creative self-efficacy and divergent thinking Exploratory Analyses: time to incubate within the control condition

4.0

Discussion

4.1

The Effects of Nature Volunteering on Creativity The Effects of Condition on Mood Inspiration and Openness to Experience Limitations

4.2 4.3 4.4 5.0

Conclusions and Recommendations

18 18 18 18 18 18 18 18

19

19 22 22 23 24 24 25

References

26

Appendix

29

List of Figures and Tables

1

Figure 1. Word cloud of the top 100 most used words in free text responses to inspiring features of the environment and activity.

Page 19

Table 1. Descriptive Statistics for the Main Variables at Pre- and Post-Test

Page 13

Table 2. Primary Outcomes: Pre-mean, Post-mean, and Mean Differences

Page 15

Table 3. Secondary Outcomes: Pre-mean, Post-mean, and Mean Differences

Page 17

Table 4. Results of Multiple Linear Regression Predicting Creative Self-Efficacy

Page 20

Table 5. Results of Multiple Linear Regression Predicting Top 2 Scores

Page 20

Table 6. Results of Multiple Linear Regression Predicting Fluency Scores

Page 20

2

Executive

Summary

3

Spending time in nature spaces is believed to provide a myriad of benefits to cognition, mood, and overall mental health and wellbeing (Hartig et al., 2014; Kaplan, 1995; Ulrich et al., 1991) with specific benefits shown for ‘blue space’ containing aquatic elements (Pedersen et al., 2019; White et al., 2010). More recently there have been efforts to explore the potential benefits of nature to creative thinking and innovation (Vella-Brodrick et al., 2024). William Wordsworth, Emily Dickinson, Claude Monet, Friedrich Nietzsche, and Rachel Carson are all examples of great thinkers or artists who sought inspiration for their work from time spent in nature (Gros, 2023; Williams et al., 2018). The Attention Restoration Theory (ART; Kaplan & Kaplan, 1989) proposes that nature spaces elicit soft fascination; that is, elements of the environment that are engaging enough to capture one’s attention slightly, but allowing enough cognitive resources for incubation, mind wandering and ‘mental housekeeping’, which may support creative thinking (Basu et al., 2019; Vella-Brodrick et al., 2024). To explore this possibility in the ‘real world’, we compared the performance of three groups of professionals on a creative thinking task before and after spending a day volunteering in nature (a wetlands reserve), non-nature settings (e.g. inside a school or care home), or conducting their ordinary day of work. The task was focused on an authentic problem – ‘ways to improve employee engagement programmes for employee and community benefit’ – with potential for suggestions that could lead to genuine corporate innovation. There were general benefits of all types of volunteering relative to controls. However, we found that nature volunteers reported the greatest increase in their sense of integration with nature and perceived their volunteering environment as the most restorative. They also showed a shift to a calmer and more positive mood state and reported feeling more inspired by their environment and activities. Benefits for creative thinking were less clear cut. Both volunteering groups showed weak evidence that their creative self-efficacy improved. Further, the nature volunteers showed some evidence of better sustaining the creative quality and quantity of ideas produced in response to the authentic problem. However, there was no evidence that volunteering in nature improved idea quality. A randomised controlled trial could enhance future evidence quality.

4

5

1.0

Introduction William Wordsworth, Emily Dickinson, Claude Monet, Friedrich Nietzsche, and Rachel Carson are all great thinkers or artists who sought inspiration for their work from time spent in nature (Gros, 2023; Williams et al., 2018). Spending time in nature spaces is believed to provide myriad benefits to one’s cognition, mood, and overall mental health and wellbeing (Hartig et al., 2014; Kaplan, 1995; Ulrich, 1991). The Attention Restoration Theory (ART; Kaplan & Kaplan, 1989) in particular, proposes that spaces that offer a sense of being away; include elements of soft fascination; are compatible with one’s goals or needs; and are perceived as coherent and having sufficient content and structure, may enact attentional restoration. Indeed, restorative environments (typically nature spaces, but not exclusively so; Collado et al., 2016) have seen an extensive body of research aimed at exploring their benefits to health. A growing number of studies suggest that there may be specific benefits of spending time in nature spaces that contain ‘blue’ elements (Seresinhe et al., 2015) and, in particular, aquatic elements such as rivers, lakes, wetlands and coasts (Barton & Pretty, 2010; Pedersen et al., 2019; White et al., 2010). However, more recently there have also been efforts to understand the benefits of nature spaces to creativity and innovation, with the hope to leverage the same benefits for idea generation reported by great thinkers throughout history (Ratcliffe et al., 2021; Vella-Brodrick et al., 2024; Williams et al., 2018).

1.1

Creativity and Innovation

Generally, creativity is regarded as a multi-stage process resulting in the production of original and effective outputs or ideas (Runco & Jaeger, 2012). Most models of the creative process (e.g., Mumford et al. 1991; Sowden et al, 2025) include a phase of exploration – identifying and framing problems to work on, retrieving and gathering relevant information, and considering different perspectives – a phase of ideation – using associative processing to generate ideas by making connections, synthesising and transforming information – and a phase of evaluation – analysing ideas generated, evaluating whether they address the problem, and selecting the best idea(s) – following which these ideas may be materialized and implemented. Sometimes ideas emerge following a period of incubation or ceasing of conscious work on a problem (Gilhooly et al, 2013). Creativity has far-reaching benefits at both an individual level, including increased confidence, self-expression, mood and well-being, and cognitive functions such as problem-solving and task switching (Acar et al., 2020; Khalil et al., 2019), and on a wider societal level (Sternberg & Karami, 2024). Further, employee creativity has been associated with innovative behaviours in organisations (Elidemir et al., 2020). Innovation typically corresponds to the generation and implementation of an idea or concept within the workplace context, resulting in the introduction of new products or services in an organization (Kahn, 2018; Zhou & Hoever, 2014), and is seen as a crucial part of competitive advantage (e.g. Elidemir et al., 2020).

6

Measurement of creativity has frequently used tasks of divergent thinking. These tasks aim to capture the free flowing, nonlinear, connective and associative ideation process, whereby multiple possibilities are generated in a short period of time (Runco & Acar, 2012; Silvia et al., 2008). A common divergent thinking task, the Alternate Uses Task (AUT), prompts participants to generate as many creative uses as possible for an object, such as a brick or a pen. Participant’s responses can then be assessed (Reiter-Palmon et al., 2019; Silvia et al., 2008) for their quantity (a proxy for fluency of thinking) and creative quality as a set (“Snapshot scoring”) or considering just the most creative responses (e.g., “Top 2” scoring), and considering the diversity or variety of ideas (a proxy for cognitive flexibility). Research has also explored divergent thinking using more real-world problems (Mumford et al., 2012) such as an original and effective way to advertise a smartphone application that rewards users for actions they take to save water (OECD, 2019). Responses to these ‘authentic’ problems, can be scored for quantity and quality of creative ideas in the same way as the AUT. Many factors are known to influence creativity, such as one’s openness to experience in everyday life (e.g., curiosity, adventurousness and open-mindedness; Feist, 1998); affect or mood (Baas et al., 2008); creative self-efficacy (e.g., the belief that one can be creative; Beghetto, 2006); and feeling inspired (Thrash et al., 2014). These factors can also extend beyond the self, for example environments impact the creative process through their effects on mood and attention, with boring, dull, or stressful environments hindering idea generation and creativity (Aristizabal et al., 2021; James et al., 2004; Zhou & Hoever, 2014).

1.2 Nature Creativity Connection Recently there has been an increase in research on using natural environments (e.g., green spaces such as forests, or blue ‘watery’ spaces such as beaches or wetlands) to

7

facilitate creative thinking (Ratcliffe et al., 2021; Vella-Brodrick et al., 2024). Nature may have a calming effect, conducive to imagination (Vesala & Tuomivaara, 2018) and some evidence suggests increased creativity at work may result from spending time in nature outside of work hours or living close by nature spaces (Brossoit et al., 2024). Theoretically, according to the ART, spending time in nature may support creativity due to its capacity to elicit soft fascination; that is, elements of the environment that are engaging enough to capture one’s attention slightly but allowing enough cognitive resources for mind wandering and ‘mental housekeeping’ (Basu et al., 2019; Vella-Brodrick et al., 2024). These elements, such as clouds, the swaying of trees in the wind, or the babble of water in a brook, are much more conducive to reflection and the incubation stage of the creative process than more intense or absorbing activities (e.g., social media, watching TV; Basu et al., 2019; Vella-Brodrick et al., 2024; Williams et al. 2018). Mind wandering may have specific benefits for divergent thinking (Rodriguez-Boerwinkle et al., 2024) by aiding unconscious associative processing during creative incubation, potentially leading to a creative breakthrough (Huang et al., 2024; Williams et al., 2018). Furthermore, Baird et al., (2012) observed that engagement in a minimally cognitively-demanding task (i.e., a choice reaction time task), similar to nature exposure, facilitated greater mind wandering, and greater creative incubation compared to a more demanding task. A further tenet of ART is that restoration may occur through a sense of ‘extent’ and ‘being away’ or “transported beyond the immediate setting” (Kaplan, 2001, p511). Tang et al. (2023) proposed that contact with nature may broaden individual’s cognitive capacity, prompting them to recall the wider world and enabling them to explore beyond the immediate context. In a series of empirical studies, they found that broader cognitive processing, as a

result of nature exposure, fosters creativity at work (see also Shalley et al., 2009). Relatedly, other work has shown that diversifying experiences, for instance exposure to other cultures (Leung & Chiu, 2010), that similarly broaden mental focus can promote creativity (Ritter et al., 2012). In summary, factors such as increasing urbanity and workplace demands necessitate exploration of the value of nature spaces as a resource to scaffold creativity and innovation but, whilst there are some promising ideas and findings, research on the connection between nature, especially blue nature spaces, and creativity is underdeveloped (e.g., Charisi et al., 2025); further research is required.

1.3

Present Research Aims and Hypotheses

The aim of the present study is to compare the impact of volunteering in wetland nature, on creative ideation and related constructs. Because volunteering itself might plausibly lead to benefits for creative ideation through social connectivity and diversifying experiences (Perry-Smith, 2006), we compared nature volunteers to a homologue condition of volunteering in a nonnature environment. Participants were professional corporate workers taking part in employee volunteering programmes either in nature or non-nature settings and control participants who completed a regular workday. Participants completed various measures and tasks before and after their volunteering day or workday, to assess their creative ideation abilities, creative selfefficacy, aspects of cognitive ability, mood, inspiration, openness, nature integration and restoration. We expected positive benefits of spending time in wetland nature over and above the benefits of volunteering in a non-nature setting, which itself should provide benefits compared to a typical day of work on all outcomes including restoration, creative ideation, cognitive ability, creative self-efficacy, and mood.

2.0

Method 2.1 Participants

2.2 Design

The participants were professional corporate workers (n = 226; 110 female, 88 male, 1 preferred not to say, 27 did not answer) from a range of organisations. The mean age of the sample was 37 years old (SD = 9.8). The range was 20 to 62 years old. The sample met the requirements of an a priori power analysis that indicated a sample size of n = 244 to detect the main effect of condition, with an effect size f = .20 (see Ma, 2006; Scott, Leritz & Mumford, 2004), α = .05, and power = .80.

Participants were allocated roughly equally to one of three conditions with n = 67 (30%) in nature volunteering activities, n = 84 (37%) in non-nature volunteering activities, and n = 75 (33%) in the control condition.

Intervention group participants were recruited through their participation in corporate volunteering programmes. Control participants were recruited through Prolific Academic and pre-selected, using an initial screening survey, as holding comparable employment to the intervention groups. They were paid for their participation at Prolific’s recommended rate.

2.3 Measures

Outliers’ inspection revealed eight cases which exceeded a Z score of +/- 3.5 on the variables valence, state openness, or fluency. We Winsorized to the next closest value. Multivariate inspection revealed three cases with a rare combination of fluency and top 2 scores. These participants had either high fluency and low top 2, or low fluency and high top 2. We decided to keep these participants considering that their pattern of response was valid and plausible.

The two intervention conditions completed a range of measures prior to and after completing their volunteering activities. The control condition completed the same measures at the beginning and end of their regular day’s work.

2.3.1 Primary Outcome Measures Three primary outcome measures focused on creativity and executive functioning. Divergent Thinking. We designed an authentic divergent thinking task, in collaboration with a large corporate employer, to encourage participants to generate solutions that were of real-world relevance for themselves and their organisation. Consequently, participants were asked to “Suggest as many ways as you can think of for corporate employee engagement (volunteering) programmes to further benefit employees and/or the community.” Participants were further instructed to “try and think of ideas that are as original and as creative as possible.” Participants’ responses were scored by two raters for fluency and creativity using both the top 2 (scored for a participant’s two most creative ideas only; Silvia et al., 2008) and

8

snapshot (scored as an overall rating of a participant’s pool of responses; Silvia et al., 2009) scoring methods. Agreement between the raters was assessed using Krippendorf’s alpha (fluency αα = .85; Top 2 αα = .87; snapshot αα = .90). A summary of the ideas can be found in the appendix, together with a link to the full set of ideas organised into 13 categories. Creative self-efficacy. We used five items adapted from Beghetto (2006, 2009), e.g. “I was good at coming up with new ideas” (1 = not true - 5 = very true). Participants were asked to answer these questions “with regard to your creativity when approaching the problem above about ideas for further benefits from volunteering programmes.” Connections Test (Salthouse et al., 2000). This test measures aspects of cognitive ability, most notably perceptual speed and general fluid intelligence (Salthouse, 2011) and was included as an objective measure of the restorativeness of different environments. Participants were presented with grids of semirandomly arranged numbers or combined numbers and letters. They were required to connect as many numbers (N; 1-2, 2-3, [...], 47-48), or numbers and letters (NL; 1-A, A-2, 2-B, [...], X-25), in sequence as possible within a 20-second time limit, per grid. They completed a number and number-letter grid at both pre and post-test timepoints. Two versions of each type of grid were created and counterbalanced to prevent exposure to identical stimuli. Scores were calculated by summing the number of correct connections for each grid and then calculating the ratio of these scores (NL / N score; Salthouse, 2011; Salthouse et al., 2008) for each timepoint. Participants are expected to typically make more correct connections for the number-only grids and therefore the ratio ought to be between 0 and 1. Participants completed practice grids prior to the real test. 2.3.2 Secondary Outcome Measures Seven secondary outcomes allowed assessment of the effect of condition on mood state, attention restoration,

9

nature integration, state openness to experience, and state inspiration. Mood. The Self-Assessment Manikin (Bradley & Lang, 1994) was used to capture three dimensions of mood: valence (happy – sad), arousal (aroused – relaxed/calm), and control (feeling of being controlled – in control). Each dimension was represented by a series of cartoon figures expressing the mood state, and participants placed a cross on the figure that they felt best described their current mood state on that dimension. There were nine possible positions for each dimension scored from 1- 9 (e.g., 1 = happy, 9 = sad). Perceived Restorativeness. The Short Form Perceived Restorativeness Scale (Negrín et al., 2017) was completed only as a post-measurement, after volunteering activities or work. Spending time in nature has been frequently found to be more restorative than other settings (e.g. Menardo et al., 2019). This scale is composed of five items which correspond to the attention restoration factors of fascination, coherence, being away, and compatibility, each scored from 1-5 (1 = doesn’t describe me well at all, 5 = describes me extremely well). For example; “this place lets me forget my everyday responsibilities, feel relaxed, and lose myself in my own thoughts” and “this is a fascinating place that keeps my curiosity alive and stops me from getting bored”. Nature Integration. The Inclusion of Nature in Self scale (INS; Schultz & Tabanico, 2007) assessed the construct of nature integration. Participants rated their connectedness to nature through an image of seven instances of two increasingly overlapping circles–one representing the self, the other representing nature (e.g., 1 = completely separate, 7 = fully-merged). State Openness to Experience. Openness to experience is often strongly related to creativity. Rather than capturing the extent to which individuals manifest the Big-5 (Goldberg, 1990) trait of openness to experience, we attempted to capture the extent to which a context or experience (e.g. volunteering in

nature) could be ‘opening’ and might therefore promote creativity. Hence, we adapted the ten openness items from the Big Five Inventory (John & Srivastava, 1999) to measure openness to experience in the moment, that is ‘state’ openness to experience. For example, “I see myself as someone who is curious about many different things” was reworded to “I feel curious about many different things”. Responses were made on a 1-5 scale (1 = disagree strongly, 5 = agree strongly). State Inspiration. We assessed state inspiration using three of the four items from the Inspiration Scale developed by Thrash and Elliot (2003). We slightly reworded the item “I experience inspiration” to “I am experiencing inspiration” to more strongly capture the state of being inspired and removed the item “something I encounter or experience inspires me”. Items are scored from 1-7 (1 = strongly disagree, 7 = strongly agree) and responses were summed to form an inspiration index. Measuring inspiration as a state captures potential motivational change resulting from a situation or experience, such as time spent in nature (Hotchin & West, 2020; Thrash & Elliot, 2003). Inspired by the Environment and Activity. Two follow-up single items asked participants to rate “To what extent did you feel inspired by the environment you spent the day working in?” and “To what extent do you feel inspired by the activity/activities you undertook today?” from 1-5 (1 = not inspired, 5 = very inspired). Participants were also asked to provide free text responses to list the particular aspects of the environment or activities that inspired them.

2.4 Procedure Participants in the intervention conditions completed the measures at their volunteering location at the beginning (pre) and end (post) of their day. Once participants arrived, they were given a brief prefatory talk regarding the voluntary activity and introduced to the research. Paper pre-questionnaire booklets were then distributed. After reading

an information sheet, participants provided informed consent before completing each of the measures in a consistent order. The first measures were the primary outcome measures starting with the measure of divergent creative thinking, then creative selfefficacy and then the connections test. Participants were instructed to work through the measures at their own pace until they reached the connections test. This test was timed and all participants in a group completed it together. Participants were given 20 seconds to complete each grid with the researcher managing the timing. There were two versions of the N and NL connections grids. One version was presented pre and one post. These were counterbalanced across participants, as was the order of completing the N & NL versions. Next participants moved onto the secondary outcome measures, which were again completed at participants’ own pace. The order for these was: state openness; SAM mood state; state inspiration; nature integration; demographic information. Participants then engaged in a volunteering activity. The naturebased volunteering condition

conducted nature conservation activities, such as coppicing, for an average of 375 minutes (6 hr’s 15 minutes). The non-nature-based volunteers worked on indoor tasks including painting a community hall or school stairwell, festive decorating or companionship in a care home, and running school workshops, for an average of 352 minutes (5 hr’s 52 minutes). At the end of the day’s volunteering activity, the participants were given a paper postquestionnaire booklet to complete. This was completed in an identical fashion to the pre booklet, with the addition of the measure of nature restoration, inserted after the nature integration measure. Both pre and post instances of the questionnaires took approximately 20 minutes to complete. After completing the post questionnaire, participants were thanked and given a debrief talk. Participants in the control condition completed online versions of the paper questionnaire booklets except for the connections test. This test requires timed physical responses and the movement time associated with drawing connections between numbers and letters using a paper and pen is different from the time to

click on, or swipe between, adjacent members of a sequence on a screen that could also differ significantly in physical size and therefore distance between items in the sequence. For these reasons we decided comparable connections test data could not be collected for the control condition. They completed the pre-version before starting their day’s work and the post-version after completing their day’s work. The mean time between completing the pre-questionnaire and starting the post questionnaire for the control condition was 649 minutes (10 hrs 49 minutes). Because of concerns about online data quality, at the end of each questionnaire the participants were asked to self-report their data quality (Aust et al., 2013). This assured participants that they would be paid for their participation regardless of their responses and then asked them to estimate the effort they had put into their answers, the level of distraction around them and, finally, whether their data should be used for analysis. No participants indicated that their data should be excluded.

10

11

3.0

Results 3.1

Plan of Analyses

First, we examine the descriptive statistics for asymmetric distribution. Next, we examine the effects of the experimental conditions on the primary outcomes – a) creative self-efficacy, b) top 2 scores, c) fluency scores, d) connections test – using Analysis of Covariance (ANCOVA) or Generalized Estimating Equations (GEE; a non-parametric equivalent of ANCOVA), if the parametric assumptions are not met. In those models, we control for state openness, valence, arousal, and perceived control. In secondary analyses, we investigate the secondary outcomes – a) arousal, b) valence, c) perceived control, d) state openness, e) state inspiration, f) nature integration, and g) perceived restorativeness – using a series of ANOVA or GEE models. These secondary models do not include covariates. We include a word cloud of the top 100 most used words from an openresponse question about what was inspiring from the environment and volunteering activity to provide qualitative insights into sources of inspiration during the day. Additionally, we run an exploratory hierarchical regression model to examine the effects of state inspiration and perceived restorativeness on the primary outcomes of creative selfefficacy, top 2 scores, and fluency scores, while controlling for state openness and nature integration. We also run a regression analysis to assess the impact of the time between pre and post measure completion (incubation time) on creativity scores.

All statistical analyses were conducted in SPSS 30.0. Missing data ranged from 1.3% to 16.6% (top 2 score). Little’s MCAR test was significant, χ²(83) = 124, p = .002, indicating that data were not missing completely at random. Missing values were therefore addressed using multiple imputations (20 imputations; combined using Rubin’s rule) under the assumption of missing at random (MAR). The results were consistent with and without multiple imputations. Consequently, we report results for the nonimputed dataset.

3.2 Descriptives The mean, standard-deviation, and skewness for the variables included in this study are shown in Table 1. As can be seen in Table 1, the connections test was highly skewed. In addition, the fluency and top 2 scores were moderately skewed. This is expected as it is easier to generate a handful of ideas, or unoriginal ones, compared to a large number of ideas, or very original ones. Because social connection has been linked to creativity (Perry-Smith, 2006), we asked participants if they had spent their day alone or with other people. Only 28 participants spent their day alone (12.4%). Among them, 24 were in the control condition, three were in the non-nature condition, and one was in the nature condition.

3.3 Manipulation Check We examined the effectiveness of the manipulation (i.e., the conditions) on the measure of perceived restorativeness, which captures fascination, coherence, being away, and compatibility with nature.

12

Table 1. Descriptive Statistics for the Main Variables at Pre- and Post-Test Variables

Min

Max

Mean

SD

Skew

Pre-measures 1.

Creative Self-Efficacy

1

5

3.03

0.91

-0.31

2.

Connections Test

0

3.21

0.57

0.44

3.76

3.

Nature Inspiration

1

7

4.18

1.34

-0.07

4.

State Inspiration

1

7

4.53

1.22

-0.37

5.

Valence

3

9

6.93

1.39

-0.42

6.

Arousal

1

8

3.71

1.78

0.10

7.

Control

1

9

5.81

1.84

0.05

8.

State Openness

1.8

4.80

3.44

0.65

-0.33

9.

Top2

1

4.50

2.07

0.77

0.77

10. Fluency

1

14

4.67

3.00

1.26

11. Creative Self-Efficacy

1

5

3.25

0.96

-0.65

12. Connections Test

0.03

1.52

0.51

0.23

0.80

13. Nature Inspiration

1

7

4.46

1.43

-0.41

14. State Inspiration

1

7

4.67

1.51

-0.70

15. Valence

3

9

7.33

1.60

-0.82

16. Arousal

1

9

3.40

2.12

0.65

17. Control

1

9

6.19

1.96

-0.55

18. State Openness

1.40

4.80

3.44

0.69

-0.54

19. Top2

1

4

1.89

0.79

0.71

20. Fluency

1

14

3.56

2.49

2.01

21. Inspired Environment

1

5

3.72

1.05

-0.75

22. Inspired Activity

1

5

3.71

1.06

-0.73

23. Perceived Restorativeness

1.20

5

3.37

0.97

-0.12

Post-measures

Univariate ANOVA with condition as a between-subject factor revealed a significant main effect of condition, F (2) = 102.56, p < .001, η2p = .32. Pairwise contrasts revealed that the nature condition was significantly higher compared to both non-nature (ΔMean = 0.93, p < .001) and control conditions (ΔMean = 1.34, p < .001). The non-nature condition was significantly higher than the control condition (ΔMean = 0.40, p = .01). Overall, these findings indicate that the manipulation was effective.

13

3.4 ANCOVA Assumptions Prior to analysis, we verified the ANCOVA assumptions for the main models, including the normality of residuals (using histogram, Q-Q plots, and Shapiro–Wilk test p > .05), homogeneity of variance across groups (using Levene test p > .05), linear relationships between covariates and the dependent variable (using scatter plots), and homoscedasticity (using White and Breush-Pagan test; p > .05).

After inspection, creative selfefficacy did not respect the normally distributed residuals assumption (Shapiro–Wilk test p < .001), the top 2 score violated the homoscedasticity assumption (Breusch-Pagan test χ²(1) = 10.5, p = .001), the fluency score violated the homogeneity of variance across groups assumption (Levene’s [df = 2] = 15.33, p < .001), and the connections test violated the homogeneity of variance across time assumption (Levene’s [df = 1] = 5.58, p = .02). Therefore, we relied on Generalized Estimating Equations (GEE) as a nonparametric equivalent of the ANCOVA. GEE extends the generalized linear model (GLM) framework to handle correlated or repeated measures data, such as pre–post designs, without requiring strict distributional assumptions (e.g., normality, sphericity). The correlation structure was specified as unrestricted, which allows for the correlation between pre and post measures to be freely estimated, and the distribution was selected based on the lowest Quasi Likelihood Model Criterion (QIC). The latter can be interpreted similarly to the Akaike information criterion (AIC) as an index of the accuracy of the model’s predictions (i.e., model fit).

3.5 Primary Outcomes Results The mean scores at pre-test, post-test and the difference between pre and post for each of the primary outcome variables can be seen in Table 2. 3.5.1 Creative Self-Efficacy GEE analysis with a Gamma log link distribution revealed a significant main effect of condition, Wald χ²(2) = 8.78, p = .012, and a significant main effect of time, Wald χ²(1) = 6.92, p = .01. The condition × time interaction was not significant, Wald χ²(2) = 4.05, p = .13. Among the covariates, state openness (Wald χ²(1) = 152.90, p < .001) and valence had significant effects (Wald χ²(1) = 3.92, p = .048). The model parameters revealed that the nature condition was associated with a non-significant 3% lower score (p = .51) compared to control, and the non-nature condition was associated with a non-significant 6% lower score (p = .16) compared to control. Time 1 and time 2 were virtually identical (p = .97). Pairwise contrasts with Bonferroni correction revealed that the main effect of condition was driven by a significantly lower score for the nonnature condition compared to the

control condition (ΔMean = -0.32, p = .01). The main effect of time was driven by significantly greater post-scores compared to pre-scores (ΔMean = 0.15, p = .01). Despite a non-significant interaction term, slope differences using ANCOVA and Bonferroni correction revealed that both the nature and non-nature condition manifested a significantly greater increase compared to control (the nature condition increased 0.20 more than control, whereas the non-nature condition increased 0.26 more than control, both p < .001). To summarize, although none of the conditions showed a statistical difference in their creative self-efficacy between time 1 and time 2, simple slope analysis suggests that both the nature and non-nature conditions manifested a significantly greater increase in comparison to the control condition. 3.5.2 Top 2 Top 2 scores were strongly correlated with snapshot creativity scores (pre scores r = .90; post scores r = .93), suggesting these measures are redundant. As top 2 better controls for participants’ fluency of responding, we conducted analysis on these

14

Table 2. Primary Outcomes: Pre-mean, Post-mean, and Mean Differences Pre-mean

Post-mean

Δ Post-pre

Nature

2.97

3.17

+0.20

Non-Nature

2.82

3.08

+0.26

Control

3.27

3.27

0

Nature

1.75

1.52

-0.23

Non-Nature

2.19

1.88

-0.31*

Control

2.19

2.22

+0.03

Nature

3.45

2.72

-0.73

Non-Nature

4.35

3.00

-1.35*

Control

7.29

5.14

-2.16*

Nature

0.55

0.49

-0.06

Non-Nature

0.53

0.52

-0.01

Creative Self-Efficacy

Top 2

Fluency

Connection Ratio

Note. *p < .05 with Bonferroni correction

scores and dropped analysis of snapshot scores. The GEE analysis with a Gamma log link distribution on top 2 revealed a significant main effect of condition, Wald χ²(2) = 23.94, p < .001, a significant main effect of time, Wald χ²(1) = 9.54, p = .002, and a significant interaction condition × time, Wald χ²(2) = 7.29, p = .03. Among the covariates, only state openness reached statistical significance (Wald χ²(1) = 11.76, p < .001). The model parameters revealed that the nature condition is associated with a 30% lower score (p < .001) compared to control, and the nonnature condition is associated with a 15% lower score (p = .01) compared to control. Time 1 and time 2 were virtually identical (p = .74). Pairwise contrasts with Bonferroni correction revealed that the main effect of condition was driven by a significantly lower score for nature condition compared to both nonnature (ΔMean = -0.39, p < .001) and control (ΔMean = -0.57, p < .001)

15

conditions. The non-nature condition was statistically equal to control (ΔMean = -0.17, p = .39). The main effect of time was driven by an overall significant decrease between pre to post (ΔMean = -0.18, p = .002). The interaction was driven by a significant decrease for the non-nature condition, from pre- to post-test (ΔMean = -0.31, p = .03). Slope differences using ANCOVA and Bonferroni correction revealed no significant differences in the rate of change between pre to post across conditions. To summarize, the non-nature condition manifested a small and significant decrease in their top 2 from pre to post. The Nature condition had overall significantly lower top 2 compared to the other two conditions but did not significantly decline from pre to post. 3.5.3 Fluency GEE analysis with a Negative Binomial log link distribution revealed a significant main effect of condition,

Wald χ²(2) = 53.56, p < .001, and a significant main effect of time, Wald χ²(1) = 55.19, p < .001. The condition × time interaction was marginally significant, Wald χ²(2) = 4.72, p = .09. Among the covariates, state openness was marginally significant (Wald χ²(1) = 3.80, p = .051) and perceived control had a significant effect (Wald χ²(1) = 5.33, p = .02).

.001). The interaction was driven by a significant decrease from pre to post for the non-nature and control conditions (respectively p < .001 and p = .002). The nature condition remained statistically the same from pre- to post-test (p = .27). Specifically, the nature condition decreased 0.62 points less than the non-nature and 1.43 points less than control.

The model parameters revealed that the nature condition was associated with a 45% lower score (p < 001) compared to control, and the nonnature condition was associated with a 42% lower score (p < .001) compared to control. Time 2 was associated with a 52% decrease compared to time 1 (p < .001).

Overall, both the non-nature and control conditions manifested a significant decrease in their fluency between pre- and post-tests, whilst the nature condition sustained their performance.

Pairwise contrasts with Bonferroni correction revealed that the main effect of condition was driven by a significantly greater score for the control condition compared to nature (ΔMean = 3.06, p < .001) and non-nature (ΔMean = 2.51, p < .001). The nature and non-nature conditions were statistically equal. The main effect of time was driven by a significant decrease from pre to post scores (ΔMean = -1.31, p <

3.5.4 Connection Test Results GEE analysis with linear distribution and identity link revealed a nonsignificant main effect of condition, Wald χ²(2) = 0.77, p = .38, a marginally significant main effect of time, Wald χ²(1) = 3.18, p = .07, and a nonsignificant interaction of condition × time, Wald χ²(2) = 0.03, p = .86. None of the covariates reached significance. In essence, both the nature and nonnature conditions remained constant between time 1 and time 2 on their executive functions as measured

by the connections test. The control condition did not complete this measure (see section 2.3.1).

3.6 Secondary Outcomes Results The mean scores at pre-test, posttest and the difference between pre and post for each of the secondary outcome variables can be seen in Table 3. 3.6.1 Nature Integration GEE analysis with Gamma log link distribution revealed a significant main effect of condition, Wald χ²(2) = 7.98, p < .02, a significant main effect of time, Wald χ²(1) = 25.99, p < .001, and a significant interaction of condition × time, Wald χ²(2) = 28.01, p < .001. The main effect of condition was driven by a significantly greater score for the nature condition compared to control (ΔMean = 0.61, p = .015). The main effect of time was driven by a significant increase from pre to post scores (ΔMean = 0.28, p < .001). The interaction was driven by a significant increase from pre to post for the nature and non-nature conditions (respectively p < .001 and

16

Table 3. Secondary Outcomes: Pre-mean, Post-mean, and Mean Differences Pre-mean

Post-mean

Δ Post-pre

Nature

4.30

5.03

+0.73*

Non-Nature

4.17

4.42

+0.25*

Control

4.08

4.00

-0.08

Nature

3.59

2.7

-0.89*

Non-Nature

3.83

3.82

-0.01

Control

3.65

3.59

-0.06

Nature

7.42

8.24

+0.82*

Non-Nature

7.09

7.53

+0.44

Control

6.33

6.32

-0.01

Nature

5.75

6.70

+0.94*

Non-Nature

5.64

5.92

+0.27

Control

6.05

6.04

-0.01

Nature

3.61

3.72

+0.11

Non-Nature

3.42

3.53

+0.11

Control

3.30

3.07

-0.22

Nature

4.84

5.37

+0.53*

Non-Nature

4.47

4.85

+0.38

Control

4.33

3.80

-0.52

Nature

n.a

4.16

Non-Nature

n.a

3.22

Control

n.a

2.82

Nature

n.a

4.33

Non-Nature

n.a

3.94

Control

n.a

2.90

Nature

n.a

4.21

Non-Nature

n.a

3.90

Control

n.a

3.04

Nature Integration

Arousal

Valence

Control

Openness

State Inspiration

Perceived Restorativeness

Inspired Environment

Inspired by Activity

Note. *p < .05 with Bonferroni correction

17

p = .01). The control group remained statistically the same (p = 1.0). Simple slopes analyses, with Bonferroni correction, suggested that the nature condition manifested a significantly greater increase compared to nonnature and control conditions. The nature condition increased 0.48 more than non-nature and 0.81 more than control, both p < .001, whereas nonnature and control were statistically equal, p > .05. 3.6.2 Arousal The GEE analysis with Gamma log link distribution revealed a marginally significant main effect of condition, Wald χ²(2) = 5.92, p = .05, a significant main effect of time, Wald χ²(1) = 6.14, p = .01, and a significant interaction of condition × time, Wald χ²(2) = 8.14, p = .02. The main effect of condition was driven by a significantly lower score for the nature condition compared to non-nature (ΔMean = -0.72, p = .04). The main effect of time was driven by a significant increase from pre to post scores (ΔMean = 0.36, p = .01). The interaction was driven by a significant decrease from pre to post for the nature condition (p = .01), while the other groups remained statistically the same (p > 0.5). 3.6.3 Valence The GEE analysis with Gamma log link distribution revealed a significant main effect of condition, Wald χ²(2) = 56.30, p < .001, a significant main effect of time, Wald χ²(1) = 15.40, p < .001, and a significant interaction condition × time, Wald χ²(2) = 9.79, p = .01. The main effect of condition was driven by a significantly greater score for the nature condition compared to the non-nature (ΔMean = 0.51, p = .02) and control conditions (ΔMean = 1.49, p < .001). The non-nature condition was also significantly higher than control (ΔMean = 0.98, p < .001). The main effect of time was driven by a significant increase from pre to post scores (ΔMean = 0.39, p < .001). The interaction was driven by a significant increase from pre to post for the nature condition (p = .01), while the other groups remained statistically the same (p > 0.5).

3.6.4 Control The GEE analysis with Gamma log link distribution revealed a non-significant main effect of condition, Wald χ²(2) = 2.86, p = .24, a significant main effect of time, Wald χ²(1) = 9.24, p = .002, and a significant interaction of condition × time, Wald χ²(2) = 7.34, p = .03. The main effect of time was driven by a significant increase from pre to post scores (ΔMean = 0.40, p = .002). The interaction was driven by a significant increase from pre to post for the nature condition (p = .01), while the other groups remained statistically the same (p > 0.5). 3.6.5 State Openness The GEE analysis with Gamma log link distribution revealed a significant main effect of condition, Wald χ²(2) = 26.60, p < .001, a non-significant main effect of time, Wald χ²(1) = 0.03, p = .85, and a significant interaction condition × time, Wald χ²(2) = 11.31, p = .01. The main effect of condition was driven by a significantly greater score for the nature condition compared to the control condition (ΔMean = 0.48, p < .001). The non-nature condition was also significantly higher than the control condition (ΔMean = 0.29, p = .01). However, pairwise comparisons with Bonferroni correction suggested that none of the conditions significantly increased or decreased their openness from pre to post (p > 0.5). 3.6.6 State Inspiration The GEE analysis with Gamma log link distribution revealed a significant main effect of condition, Wald χ²(2) = 26.22, p < .001, no significant main effect of time, Wald χ²(1) = 0.80, p = .37, but a significant interaction condition × time, Wald χ²(2) = 18.61, p < .001. The main effect of condition was driven by a significantly greater score for the nature condition compared to non-nature (ΔMean = 0.43, p = .02) and control conditions (ΔMean = 1.04, p < .001). The non-nature condition was also significantly higher than control (ΔMean = 0.60, p = .003). The interaction was driven by a significant increase from pre to post for the nature condition (p < .001), while the

other groups remained statistically the same (p > 0.5). 3.6.7 Perceived Restorativeness The GEE analysis with Gamma log link distribution revealed a significant main effect of condition, Wald χ²(2) =116.90, p < .001. Because none of the conditions completed the perceived restorativeness scale at pre-test, the effect of time could not be estimated. The nature condition was significantly higher than both the non-nature (ΔMean = 0.93, p < .001) and control conditions (ΔMean = 1.34, p < .001). The non-nature condition was also significantly higher than the control condition (ΔMean = 0.40, p = .01). 3.6.8 Inspired by the Environment The GEE analysis with Gamma log link distribution revealed a significant main effect of condition, Wald χ²(2) =71.96, p < .001. Because none of the conditions completed the questions at pre-test, the effect of time could not be estimated. The nature condition was significantly higher compared to both non-nature (ΔMean = 0.39, p = .004) and control conditions (ΔMean = 1.43, p < .001). The non-nature condition was also significantly higher than the control condition (ΔMean = 1.04, p < .001). Sources of inspiration in both the environment and activity, reported by participants, can be seen in Figure 1. 3.6.9 Inspired by the Activity The GEE analysis with Gamma log link distribution revealed a significant main effect of condition, Wald χ²(2) =42.36, p < .001. Because none of the conditions completed the questions at pre-test, the effect of time could not be estimated. Both the nature and non-nature conditions were significantly higher than the control condition (respectively, ΔMean = 1.17, p < .001 and ΔMean = 0.86, p < .001). The nature and non-nature conditions were statistically equal (ΔMean = 0.31, p = .07).

18

Figure 1. Word cloud of the top 100 most used words in free text responses to inspiring features of the environment and activity. Words that simply repeated elements of the prompt question (e.g. ‘inspired’) were excluded. imaginative focus contact

experience tasks pleasant

achieved together positive results

beautiful

space

spend

spent school helping great better trees beauty close

local physical birds productive thinking service staff

helped ideas

curious

manual

nature community clean users

place everyone creative children finish weather

teamwork hearing fresh creativity happy little cutting colleagues

interested

contribute

meeting difference lovely towards building purpose think stories energy knowledge faces others students pasted thoughts volunteering friends

interesting creating contributing calmness

3.7 Exploratory Regression Model: effects of inspiration and restoration on creative self-efficacy and divergent thinking We examined the contribution of state inspiration and perceived restorativeness towards the primary outcomes of creative self-efficacy, top 2 scores, and fluency scores, while controlling in a first step for state openness and nature integration. These analyses disregarded participants’ allocation to volunteering conditions or control. The results for creative self-efficacy, presented in Table 4, suggested that state inspiration and state openness both uniquely contributed to greater creative self-efficacy at post-test (Total Model F (4, 216) = 47.25, p < .001). The results for top 2 scores, presented in Table 5, suggested that perceived restorativeness was marginally significantly predicting lower top 2 scores (Total Model F (4, 216) = 1.74, p = .14). Lastly, the results

19

feeling

wildlife outside making enjoy interest create thought light illegible motivated bring giving outdoors clients getting natural learn impact painting

passion

centre

connecting encouragement

makes everything therapeutic

for fluency scores, presented in Table 6, suggested that none of the predictors contributed significantly to explaining inter-individual variance in those scores (Total Model F (4, 216) = 1.01, p = .36).

3.8 Exploratory Analyses: time to incubate within the control condition The average time elapsed between pre and post, for the control group, was 10h48 (ranged from 6h49 to 16h05). Z-scores revealed no outliers (±Z = 3.5). The results of a linear regression predicting top 2 and fluency scores from completion time, while controlling for the pre scores of top 2 and fluency in their respective model, suggested that completion time was a significant predictor of top 2 (β = -.36, p = .02), but not a significant predictor of fluency scores (β = .02, p = .89).

Table 4. Results of Multiple Linear Regression Predicting Creative Self-Efficacy Beta unstandardized

SE

ββeta standardized

t

p

State Openness

0.88

0.08

.63

11.52

.00

Nature Integration

0.02

0.04

.02

0.39

.69

State Openness

0.55

0.10

.40

5.44

.00

Nature Integration

-0.00

0.04

-.01

-0.09

.93

State Inspiration

0.19

0.05

.30

4.00

.00

Perceived Restoration

0.06

0.06

.07

1.06

.29

Total

R2adj = 45.7%

Variables Step 1

Step 2

.00

Note. n = 226 Table 5. Results of Multiple Linear Regression Predicting Top 2 Scores Beta unstandardized

SE

ββeta standardized

t

p

State Openness

0.06

0.09

.05

0.62

.54

Nature Integration

-0.05

0.04

-.08

-1.00

.32

State Openness

0.19

0.12

.16

1.55

.12

State Inspiration

-0.03

0.04

-.05

-0.60

.55

Nature Integration

-0.03

0.06

-.05

-0.48

.64

Perceived Restoration

-0.15

0.07

-.18

-2.01

.05

Total

R2adj = 1.5%

Variables Step 1

Step 2

.14

Note. n = 226 Table 6. Results of Multiple Linear Regression Predicting Fluency Scores Beta unstandardized

SE

βeta standardized

t

p

State Openness

0.23

0.28

.06

0.83

.41

Nature Integration

-0.05

0.14

-.03

-0.36

.72

State Openness

0.55

0.38

.15

1.47

.14

State Inspiration

-0.01

0.14

-.01

-0.06

.95

Nature Integration

-0.07

0.18

-.04

-0.39

.70

Perceived Restoration

-0.364

0.223

-.14

-1.63

.10

Total

R2adj = 0.2%

Variables Step 1

Step 2

.36

Note. n = 226

20

21

4.0

Discussion Using an ecologically valid task design, we measured the effect of volunteering in nature spaces on creative self-efficacy and creative thinking using a divergent thinking task that presented participants with an authentic problem (ways to improve employee engagement programmes). Participants either spent their day volunteering in nature (a wetlands reserve), volunteering in an indoor nonnature setting (a school, care home or community hall) or completed a normal workday. We expected positive benefits of spending time volunteering in nature over and above volunteering in a nonnature setting (Perry-Smith, 2006), which would prove beneficial compared to a normal day of work. The main results point towards very small benefits of both nature and non-nature volunteering on participants’ creative self-efficacy. Further, whilst both volunteering conditions showed a reduction in idea quality at post-test, this was only significant for the nonnature volunteers. In addition, non-nature volunteers and controls both showed significant reductions in the fluency of idea generation at post-test. Put another way, nature volunteers better maintained their quantity and quality of idea generation over the course of their day. However, there was no evidence that volunteering in nature increased idea quality. These results applied to participants with equal levels of state openness, arousal, valence, and perceived control (i.e., the models controlled for these cofounding factors).

4.1

The Effects of Nature Volunteering on Creativity

Our analyses supported that, as expected, volunteering in wetland nature led to higher reported levels of perceived restoration and nature integration than for nonnature volunteers and a control condition. These findings agree with other research showing benefits of blue nature spaces for restoration and wellbeing (Barton & Pretty, 2010; Pedersen et al., 2019; Seresinhe et al., 2015; White et al., 2010). However, none of the conditions differed significantly on the Connections Test, a general measure of cognitive ability, suggesting that differences in perceived restoration were not mirrored by a more objective measure of attention restoration. Based on prior theory (Williams et al., 2018) and evidence (Ratcliffe et al., 2021; Vella-Brodrick et al., 2024), we expected that the increased restoration experienced by the nature-based volunteers would translate into benefits for creative thinking. However, there was only weak evidence for these benefits. Both volunteering intervention conditions showed an increase in self-reported creative self-efficacy (CSE) from pre to post, whereas the control condition showed no change. There was no measurable specific benefit of volunteering in nature. One possible explanation is that participating in any volunteering activity enhances mood, which has been shown to increase CSE scores (He & Wong, 2022). However, only the nature

22

volunteers showed an increase in mood from pre to post and differences in mood state between conditions were statistically controlled for in our analyses. Another possible explanation is that whilst volunteering in nature did benefit CSE as predicted, the nonnature volunteering tasks in this study, which involved ‘creative activities’ including painting and Christmas decorating, may also have served to enhance participants’ sense of CSE, leading to no-difference between volunteering conditions and to both showing evidence of increased CSE relative to controls. Turning to our objective measures of divergent creative thinking, surprisingly, the rated creativity of the top 2 ideas produced to the authentic problem was lower for the nature volunteers compared to the other conditions both before and after the volunteering activity. Further, on both the pre and post measures, the nature and non-nature volunteers produced fewer ideas than the controls. As differences between conditions were present at pre-test this implies that they were pre-existing differences between the groups of participants. For instance, these may have arisen

23

because expectations about the volunteering day ahead ‘primed’ participants’ idea generation process even at pre-test. Nature volunteers, as a group, were more likely to produce employee engagement ideas related to conservation and nature, yielding a relatively homogenous (less creative) set of ideas compared to non-nature volunteers (and controls), who engaged in range of different volunteering activities (or days of work) that primed a more diverse and creative set of ideas. Other differences between conditions in the experiences participants had before they started the pre-test measures (e.g. whether they were working with others or alone; levels of distraction) may have similarly contributed to baseline differences. Whilst we are not able to definitively explain baseline differences in creativity scores with the current measures, because our design compares change from pre to post, we are still able to measure the impact of the different conditions over time irrespective of the pre-existing between-group differences. Notably, idea quality significantly decreased over time for the nonnature volunteers, whereas this was

not true for the nature volunteers or controls, and idea quantity decreased significantly over time for both the non-nature volunteers and controls. Because all groups produced ideas to the same authentic problem at pre and post test we would expect some reduction in the number of new ideas at post-test, as participants are likely to generate the majority of their responses when they are first presented with the problem at pre-test (see also Baird et al, 2012). Viewed in this context, the maintenance of the quantity of ideas produced over time by the nature volunteers points to a benefit of volunteering in nature for sustaining the quantity of production of ideas. In addition, the non-significant change in idea quality over time for the nature volunteers suggests that relative to other types of volunteering activity, but not controls, there may be a small benefit of volunteering in nature for sustaining the quality of production of ideas.

4.2 The Effects of Condition on Mood Both volunteering conditions showed significantly more positively

valenced mood state at pre-test, perhaps because they were looking forwards to their day volunteering in comparison to the regular workday ahead of the controls. However, the nature volunteers showed a further significant increase in valence (see also Reeves et al., 2019), as well as a stronger sense of being in control at post-test, alongside a decrease in arousal (i.e. they were more relaxed). Taken together, this is clear evidence that volunteering in nature had benefits for mood. However, this may have somewhat acted as a doubleedged sword, partially explaining why the benefit of volunteering in nature on creativity was not stronger. Whilst previous research has shown a relationship between mood and creativity, positive mood only predicts creativity when it is an activating mood state, such as joy or excitement, rather than a deactivating mood state, such as relaxation (Baas et al. 2008), which was the mood state the nature volunteers moved towards.

4.3 Inspiration and Openness to Experience We measured inspiration using a state inspiration scale developed by Thrash and Elliot (2003) and individual items asking volunteers the extent to which they were inspired by their volunteering activity and environment. Overall, the nature volunteers reported a greater sense of being in an inspired state and were more inspired by their environment than non-nature volunteers or controls. Both volunteering conditions were more inspired by their activity than controls. The increase in inspiration for the nature condition was not linked to any change in openness to experience, which has previously been linked to both inspiration (Thrash & Elliot, 2003) and creativity (Feist, 1998). i.e. there was no evidence that a nature environment or volunteering activities were more ‘opening’ despite their being more inspiring. Whilst feeling inspired predicted participants having an increased sense of creative self-efficacy there were no effects of inspiration on objectively measured creative divergent thinking, in contrast to

previous reports (Thrash et al, 2014). Similarly to state inspiration, a sense of openness also predicted CSE but there was no effect on the fluency of ideas or top 2 scores of idea creativity. Overall, our findings suggest that nature has benefits more for participants’ self-perceptions than for objective task performance. Nevertheless, research has indicated that creative self-perceptions can promote engagement with creative activities and creative achievement (Beghetto & Karwowski, 2025), but these effects are only likely to emerge over a longer time frame than the present study.

4.4 Limitations Research suggests that having greater incubation time results in the generation of more ideas (Silveira, 1971). Due to the ecologically valid conditions of our study, participants’ incubation time (the time between pre and post-test) naturally varied across conditions as a function of the volunteering activity or day of work, which could have affected the quality and quantity of ideas produced at post-test. In particular, whilst the nature and non-nature volunteers spent similar amounts of time volunteering (averaging 375 minutes and 352 minutes respectively) the control group spent far longer engaged in their day of work (649 minutes on average). The control group also produced more ideas on average than the volunteering groups and better quality ideas than the nature group. In the present study we were unable to include incubation time as a co-variate to control for this potential effect because there was no variation at the individual level for participants in the volunteering conditions. All volunteers undertook the same activity for the same length of time as other members in their group. Therefore, we explored the possible relationship between incubation time and post-test creativity, using regression analysis on the control condition participants to test if differences in incubation time affected the quantity and quality of idea generation. We found that incubation time did not predict idea fluency and negatively predicted

idea quality. More time to incubate was associated with poorer ideas, making it less likely that differences in incubation time explain differences in creativity on the authentic task used in the present study. A further limitation of our design is that participants were allocated to conditions as a result of preexisting relationships between their respective employer and the volunteering organisation or by their pre-existing use of the Prolific Academic platform. To try and counter this, participants were matched across conditions on type of employment and professional status. Nevertheless, it is possible that unmeasured systematic differences between the participants in each condition may have influenced the results. This possibility is evident from baseline differences between groups on some of our measures. Future work should use random allocation of participants to condition.

24

5.0

Conclusions and Recommendations The main aim of the present research was to explore whether spending time in wetland nature resulted in benefits for creative thinking, which is a key requirement for corporate innovation. Our findings suggest that volunteering in general may benefit creative self-efficacy, and that volunteering in nature benefitted inspiration. Both of these factors are known to be important antecedents of engagement in creative activities. Further, there was some evidence that volunteering in nature may have a small impact on the maintenance of creative thinking over the course of a day. This may reflect the greater restorative benefits of spending time in nature. Our findings clearly show that spending time in wetland nature resulted in higher perceived restoration and a shift towards a more positive and relaxed mood state, pointing to potential wellbeing benefits of wetland nature. The interpretation of our findings is complicated by baseline differences across conditions in quality and quantity of creative ideas and differences in incubation time over the day. These differences are likely the result of the ecologically valid approach used in the present study, which sought to assess the impact of nature and non-nature volunteering as implemented through corporate employee engagement programmes. However, as a result, there may have been systematic differences between conditions that impacted findings, despite attempts to match participants on occupational status. Nevertheless, in sum, the present research showed clear benefits of volunteering on a range of measures, with specific benefits of volunteering in wetland nature on mental restoration, inspiration and mood

25

and evidence of some modest effects on creativity.

Going forwards it is recommended to: 1. Strengthen and Expand Nature Based Volunteering Opportunities Because wetland-based volunteering consistently produced higher perceived restoration, greater inspiration, and more positive mood, organisations should increase the availability and visibility of naturebased volunteering days. These activities appear particularly effective for supporting wellbeing and may help sustain creative output over the course of a day, even if they do not directly increase creative quality. 2. Consider Timing of Volunteering Activity to Leverage Benefits for Idea Generation Creative problem-solving often benefits from time-away from the problem to incubate ideas whilst taking part in different activities that are engaging but not overly demanding. Placing nature-based volunteering in the middle of working on a problem requiring creative ideas, for instance part-way through a workshop, offers potential benefits for idea incubation. 3. Implement a Randomised Controlled Trial (RCT) to Establish Causal Effects Randomly allocating participants to nature based volunteering, non-nature volunteering, or control conditions, together with standardisation across conditions of volunteering time, social contact and range of activities, would remove the impact of potential systematic differences between conditions on research findings.

4. Incorporate Longitudinal Follow Up to Measure Downstream Impacts Because creativity related benefits such as creative self-efficacy and inspiration may exert their influence over longer periods, future evaluations should include follow up measurements beyond the volunteering day (e.g., 1 week, 1 month, and 3 months). Tracking outcomes over time would help determine whether short term boosts in mood, restoration, and inspiration translate into sustained engagement in creative activities, higher-quality idea generation, or increased innovative behaviour in workplace settings. Such longitudinal data would also enable organisations to understand the longer-term return on investment of naturebased volunteering programmes for employee development and wellbeing. 5. Use Ecological Momentary Assessment (EMA) to Capture Real Time Dynamics To better understand how inspiration, mood, restoration, and creative cognition fluctuate throughout the volunteering day, future research should employ ecological momentary assessment (EMA) methods. By prompting participants at multiple random points during the day, EMA would capture real-time changes in key variables rather than relying on pre- and post-event snapshots. This approach would offer richer insights into when and how nature-based activities influence creative processes, such as whether benefits occur during the activity itself, during downtime, or in reflective moments afterwards. EMA could also help disentangle the influence of fatigue, social interaction, environmental stimulation, and task demands on creativity related outcomes..

References

Acar, S., Tadik, H., Myers, D., Sman, C., & Uysal, R. (2020). Creativity and Well‐being: A Meta‐analysis. The Journal of Creative Behavior, 55(3), 738–751. https://doi. org/10.1002/jocb.485 Aristizabal, S., Byun, K., Porter, P., Clements, N., Campanella, C., Li, L., Mullan, A., Ly, S., Senerat, A., Nenadic, I. Z., Browning, W. D., Loftness, V., & Bauer, B. (2021). Biophilic office design: Exploring the impact of a multisensory approach on human well-being. Journal of Environmental Psychology, 77, 101682. https://doi.org/10.1016/j. jenvp.2021.101682 Aust F., Diedenhofen B., Ullrich S., & Musch J. (2013). Seriousness checks are useful to improve data validity in online research. Behavior Research Methods. 45(2), 527-35. https://doi. org/10.3758/s13428-012-0265-2

Basu, A., Duvall, J., & Kaplan, R. (2019). Attention Restoration Theory: Exploring the role of soft fascination and Mental Bandwidth. Environment and Behavior, 51(9-10), 1055–1081. https://doi. org/10.1177/0013916518774400 Beghetto, R. A. (2006). Creative Self-Efficacy: Correlates in Middle and Secondary Students. Creativity Research Journal, 18(4), 447–457. https://doi.org/10.1207/ s15326934crj1804_4 Beghetto, R. A. (2009). Correlates of intellectual risk taking in elementary school science. Journal of Research in Science Teaching, 46(2), 210–223. https://doi.org/10.1002/tea.20270 Beghetto, R. A., & Karwowski, M. (2025). Creative Agency Unbound. Cambridge University Press.

Baas, M., De Dreu, C. K. W., & Nijstad, B. A. (2008). A meta-analysis of 25 years of mood-creativity research: Hedonic tone, activation, or regulatory focus? Psychological Bulletin, 134(6), 779–806. https://doi. org/10.1037/a0012815

Bradley, M. M., & Lang, P. J. (1994). Measuring emotion: The selfassessment manikin and the semantic differential. Journal of Behavior Therapy and Experimental Psychiatry, 25(1), 49– 59. https://doi.org/10.1016/00057916(94)90063-9

Baird, B., Smallwood, J., Mrazek, M. D., Kam, J. W. Y., Franklin, M. S., & Schooler, J. W. (2012). Inspired by Distraction: Mind Wandering Facilitates Creative Incubation: Mind Wandering Facilitates Creative Incubation. Psychological Science, 23(10), 1117-1122. https://doi. org/10.1177/0956797612446024

Brossoit, R. M., Crain, T. L., Leslie, J. J., Fisher, G. G., & Eakman, A. M. (2024). Engaging with nature and work: associations among the built and natural environment, experiences outside, and job engagement and creativity. Frontiers in Psychology, 14, 1268962. https://doi.org/10.3389/ fpsyg.2023.1268962

Barton J. & Pretty J. (2010). What is the best dose of nature and green exercise for improving mental health? A multi-study analysis. Environmental Science & Technology.44(10), 3947-55. https://doi.org/10.1021/es903183r

Charisi, V., Zafeiroudi, A., Trigonis, I., Tsartsapakis, I., & Kouthouris, C. (2025). The Impact of Green Spaces on Workplace Creativity: A Systematic Review of Nature-Based Activities and Employee Well-Being.

Sustainability, 17(2), 390. https://doi. org/10.3390/su17020390 Collado, S., Staats, H., Corraliza, J. A., & Hartig, T. (2016). Restorative environments and health. In Handbook of Environmental Psychology and Quality of Life Research (pp. 127148). Cham: Springer International Publishing. Elidemir, S. N., Ozturen, A., & Bayighomog, S. W. (2020). Innovative Behaviors, Employee Creativity, and Sustainable Competitive Advantage: A Moderated Mediation. Sustainability, 12(8), 3295. https://doi.org/10.3390/ su12083295 Feist, G. J. (1998). A Meta-Analysis of Personality in Scientific and Artistic Creativity. Personality and Social Psychology Review, 2(4), 290–309. https://doi.org/10.1207/ s15327957pspr0204_5 Gilhooly, K. J., Georgiou, G., & Devery, U. (2013). Incubation and creativity: Do something different. Thinking & Reasoning, 19(2), 137–149. https://doi. org/10.1080/13546783.2012.749812 Goldberg, L. R. (1990). An alternative” description of personality”: the big-five factor structure. Journal of Personality and Social Psychology, 59(6), 1216-1229. https://doi. org/10.1037//0022-3514.59.6.1216 Gros, F. (2023). A Philosophy of Walking. Verso Books. Hartig, T., Mitchell, R., de Vries, S., & Frumkin, H. (2014). Nature and Health. Annual Review of Public Health, 35(1), 207–228. https://doi.org/10.1146/ annurev-publhealth-032013-182443

26

He, W.-j., & Wong, W.-c. (2022). Affective state contributes to creative self-efficacy: Evidence from an experimental study of emotion induction. Thinking Skills and Creativity, 45, 1–11. https://doi.org/10.1016/j. tsc.2022.101061 Hotchin, V., & West, K. (2020). Reflecting on nostalgic, positive, and novel experiences increases state Openness. Journal of Personality, 89(2), 258–275. https://doi.org/10.1111/ jopy.12580 Huang, Y., Song, X., & Ye, Q. (2024). Mind wandering and the incubation effect: Investigating the influence of working memory capacity and cognitive load on divergent thinking. Thinking Skills and Creativity, 52, 101499. https://doi.org/10.1016/j. tsc.2024.101499 James, K., Brodersen, M., & Eisenberg, J. (2004). Workplace Affect and Workplace Creativity: A Review and Preliminary Model. Human Performance, 17(2), 169– 194. https://doi.org/10.1207/ s15327043hup1702_3 John, O. P., & Srivastava, S. (1999). The Big Five Trait taxonomy: History, measurement, and theoretical perspectives. In L. A. Pervin & O. P. John (Eds.), Handbook of personality: Theory and research (pp. 102–138). Guilford Press. Kahn, K. B. (2018). Understanding innovation. Business Horizons, 61(3), 453–460. https://doi.org/10.1016/j. bushor.2018.01.011 Kaplan R. (2001). The nature of the view from home: Psychological benefits. Environment and Behavior, 33, 507-542. https://doi. org/10.1177/00139160121973115 Kaplan, S. (1995). The restorative benefits of nature: Toward an integrative framework. Journal of Environmental Psychology, 15(3), 169182. https://doi.org/10.1016/02724944(95)90001-2 Kaplan, R., & Kaplan, S. (1989). The experience of nature: A psychological perspective. Cambridge University Press.

27

Khalil, R., Godde, B., & Karim, A. A. (2019). The Link Between Creativity, Cognition, and Creative Drives and Underlying Neural Mechanisms. Frontiers in Neural Circuits, 13. https:// doi.org/10.3389/fncir.2019.00018 Leung, A. K., & Chiu, C. (2010). Multicultural Experience, Idea Receptiveness, and Creativity. Journal of Cross-Cultural Psychology, 41(5-6), 723-741. https://doi. org/10.1177/0022022110361707 Ma, H. H. (2006). A synthetic analysis of the effectiveness of single components and packages in creativity training programs. Creativity Research Journal, 18(4), 435-446. https://doi.org/10.1207/ s15326934crj1804_3 Menardo, E., Brondino, M., Hall, R., & Pasini, M. (2019). Restorativeness in Natural and Urban Environments: A Meta-Analysis. Psychological Reports, 124(2), 417-437. https://doi. org/10.1177/0033294119884063 Mumford, M. D., Medeiros, K. E., & Partlow, P. J. (2012). Creative thinking: Processes, strategies, and knowledge. The Journal of Creative Behavior, 46(1), 30–47. https://doi. org/10.1002/jocb.003 Mumford, M. D., M. I. Mobley, R. Reiter-Palmon, C. E. Uhlman, and L. M. Doares. (1991). Process Analytic Models of Creative Capacities. Creativity Research Journal, 4(2), 91–122. https://doi. org/10.1080/10400419109534380 Negrín, F., Hernández-Fernaud, E., Hess, S., & Hernández, B. (2017). Discrimination of Urban Spaces with Different Level of Restorativeness Based on the Original and on a Shorter Version of Hartig et al.’s Perceived Restorativeness Scale. Frontiers in Psychology, 8. https://doi. org/10.3389/fpsyg.2017.01735 OECD. (2019). Framework for the Assessment of Creative Thinking in PISA2021: Third Draft. OECD Publishing. https://www.oecd.org/ content/dam/oecd/en/topics/ polic y-sub-issues/creat ive-think ing/ PISA-2021-creative-think ingframework.pdf

Pedersen E., Weisner S. E. B., Johansson M. (2019). Wetland areas’ direct contributions to residents’ well-being entitle them to high cultural ecosystem values. Science of the Total Environment.646, 3151326. https://doi.org/10.1016/j. scitotenv.2018.07.236 Perry-Smith, J. E. (2006). Social Yet Creative: The role of social relationships in facilitating individual creativity. Academy of Management Journal, 49(1), 85–101. https://doi. org/10.5465/AMJ.2006.20785503 Ratcliffe, E., Gatersleben, B., Sowden, P. T., & Korpela, K. M. (2021). Understanding the Perceived Benefits of Nature for Creativity. The Journal of Creative Behavior, 56(2), 215–231. https://doi.org/10.1002/jocb.525 Reeves J. P., Knight A. T., Strong E. A., Heng V., Neale C., Cromie R., & Vercammen A. (2019). The Application of Wearable Technology to Quantify Health and Wellbeing Co-benefits From Urban Wetlands. Frontiers in Psychology. 10,1840. https://doi. org/10.3389/fpsyg.2019.01840 Reiter-Palmon, R., Forthmann, B., & Barbot, B. (2019). Scoring divergent thinking tests: A review and systematic framework. Psychology of Aesthetics, Creativity, and the Arts, 13(2), 144–152. https://doi.org/10.1037/ aca0000227 Ritter, S. M., Damian, R. I., Simonton, D. K., van Baaren, R. B., Strick, M., Derks, J., & Dijksterhuis, A. (2012). Diversifying experiences enhance cognitive flexibility. Journal of Experimental Social Psychology, 48(4), 961–964. https://doi.org/10.1016/j. jesp.2012.02.009 Rodriguez-Boerwinkle, R. M., Welhaf, M. S., Smeekens, B. A., Booth, R. A., Kwapil, T. R., Silvia, P. J., & Kane, M. J. (2024). Variation in divergent thinking, executive-control abilities, and mindwandering measured in and out of the laboratory. Creativity Research Journal, 1-35. https://doi.org/10.1080/1040 0419.2024.2326336 Runco, M. A., & Acar, S. (2012). Divergent thinking as an indicator of creative potential. Creativity Research

Journal, 24(1), 66-75. https://doi.org/1 0.1080/10400419.2012.652929

doctoral dissertation, University of Oregon.

201-230. https://doi.org/10.1016/ S0272-4944(05)80184-7

Runco, M. A., & Jaeger, G. J. (2012). The standard definition of creativity. Creativity Research Journal, 24(1), 92–96. https://doi.org/10.1080 /10400419.2012.650092

Silvia, P. J., Winterstein, B. P., Willse, J. T., Barona, C. M., Cram, J. T., Hess, K. I., ... & Richard, C. A. (2008). Assessing creativity with divergent thinking tasks: exploring the reliability and validity of new subjective scoring methods. Psychology of Aesthetics, Creativity, and the Arts, 2(2), 68. https://doi. org/10.1037/1931-3896.2.2.68

Vella-Brodrick, D. A., Lewis, K. J., & Gilowska, K. (2024). Exploring the Nature-Creativity Connection Across Different Settings: A Scoping Review. Educational Psychology Review, 36(4), 134. https://doi.org/10.1007/ s10648-024-09964-0

Salthouse, T. A. (2011). What cognitive abilities are involved in trail-making performance? Intelligence, 39(4), 222-232. https://doi.org/10.1016/j. intell.2011.03.001 Salthouse, T. A., Pink, J. E., & TuckerDrob, E. M. (2008). Contextual analysis of fluid intelligence. Intelligence, 36(5), 464–486. https://doi.org/10.1016/j. intell.2007.10.003 Salthouse, T. A., Toth, J., Daniels, K., Parks, C., Pak, R., Wolbrette, M., & Hocking, K. J. (2000). Effects of aging on efficiency of task switching in a variant of the Trail Making Test. Neuropsychology, 14(1), 102–111. https://doi.org/10.1037/08944105.14.1.102 Schultz, P. W., & Tabanico, J. (2007). Self, Identity, and the Natural Environment: Exploring Implicit Connections With Nature. Journal of Applied Social Psychology, 37(6), 1219–1247. https://doi.org/10.1111/ j.1559-1816.2007.00210.x Scott, G., Leritz, L. E., & Mumford, M. D. (2004). The Effectiveness of Creativity Training: A Quantitative Review. Creativity Research Journal, 16(4), 361–388. https://doi. org/10.1207/s15326934crj1604_1 Seresinhe, C. I., Preis, T. & Moat, H. S. (2015). Quantifying the impact of scenic environments on health. Scientific Reports, 5. 16899. https:// doi.org/10.1038/srep16899 Shalley C. E., Gilson L. L., & Blum T. C. (2009). Interactive effects of growth need strength, work context, and job complexity on self-reported creative performance. Academy of Management Journal, 52, 489505. https://doi.org/10.5465/ AMJ.2009.41330806 Silveira, J. M. (1971). Incubation: The effect of interruption timing and length on problem solution and quality of problem processing. Unpublished

Silvia, P. J., Martin, C., & Nusbaum, E. C. (2009). A snapshot of creativity: Evaluating a quick and simple method for assessing divergent thinking. Thinking Skills and Creativity, 4(2), 79-85. https://doi.org/10.1016/j. tsc.2009.06.005 Sowden, P. T., Warren, F., Seymour, M., Martin, C. RF, Kauer, A. RF, Spencer, E. RF, Mansfield, S. RF, & Waite, J. (2025). A Creativity Navigator to Guide Teaching for Creativity: Implementation and Teacher Impacts in a Creativity Collaborative of Schools. Journal of Creative Behavior, 59, e70005. https://doi.org/10.1002/ jocb.70005 Sternberg, R. J., & Karami, S. (2024). Transformational Creativity. Palgrave MacMillan. Tang, P. M., Klotz, A., McClean, S., & Lee, R. (2024). From natural to novel: The cognition-broadening effects of contact with nature at work on creativity. Journal of Management, 50(7), 2490-2533. https://doi. org/10.1177/01492063231172182 Thrash, T. M., & Elliot, A. J. (2003). Inspiration as a psychological construct. Journal of Personality and Social Psychology, 84(4), 871–889. https://doi.org/10.1037/00223514.84.4.871

Vesala, H., & Tuomivaara, S. (2018). Experimenting with work practices in a liminal space: A working period in a rural archipelago. Human Relations, 71(10), 1371-1394. https://doi. org/10.1177/0018726717744034 White, M., Smith, A., Humphryes, K., Pahl, S., Snelling, D., & Depledge, M. (2010). Blue space: The importance of water for preference, affect, and restorativeness ratings of natural and built scenes. Journal of Environmental Psychology, 30(4), 482–493. https:// doi.org/10.1016/j.jenvp.2010.04.004 Williams, K. J., Lee, K. E., Hartig, T., Sargent, L. D., Williams, N. S., & Johnson, K. A. (2018). Conceptualising creativity benefits of nature experience: Attention restoration and mind wandering as complementary processes. Journal of Environmental Psychology, 59, 36–45. https://doi. org/10.1016/j.jenvp.2018.08.005 Zhou, J., & Hoever, I. J. (2014). Research on Workplace Creativity: A Review and Redirection. Annual Review of Organizational Psychology and Organizational Behavior, 1(1), 333–359. https://doi.org/10.1146/ annurev-orgpsych-031413-091226

Thrash, T. M., Moldovan, E. G., Oleynick, V. C., & Maruskin, L. A. (2014). The psychology of inspiration. Social and Personality Psychology Compass, 8(9), 495–510. https://doi. org/10.1111/spc3.12127 Ulrich, R. S., Simons, R. F., Losito, B. D., Fiorito, E., Miles, M. A., & Zelson, M. (1991). Stress recovery during exposure to natural and urban environments. Journal of Environmental Psychology, 11(3),

28

Appendix Volunteering Ideas Participants were asked to “Suggest as many ways as you can think of for corporate employee engagement (volunteering) programmes to further benefit employees and/or the community.” Participants were further instructed to “Try and think of ideas that are as original and as creative as possible.” In total participants produced 1,678 ideas (1,363 after removing near duplicates). These ideas were grouped into 13 categories through interactive prompting with M365 Copilot and manual allocation to categories. The full set of de-duplicated verbatim ideas produced by participants, organised into these categories can be viewed here: https://osf.io/ecydh/files/osfstorage/699b5f90a4fbe80c54fef55a

Categories of Ideas 1.

Programme Design & Employee Engagement Strategies (e.g., paid volunteering days, newsletters/updates, recognition/awards, family involvement, internal portal, manager support, competitions/challenges, flexible formats).

2. Environmental Conservation & Nature (e.g., tree planting, litter/beach/river cleanups, habitat/wildlife support, parks/gardens, climate/sustainability education). 3. Education, Schools & Youth Development (e.g., mentoring, reading in schools, STEM days, careers/CV/interview support, work experience/shadowing, university/ college links). 4. Homelessness, Food Poverty & Donation Drives (e.g., soup kitchens, food banks, clothing drives, fundraising events). 5. Elderly, Loneliness & Care Settings (e.g., visiting care homes, befriending/ loneliness initiatives, dementia support). 6. Health, NHS & Public Health (e.g., NHS hospital roles, blood donation, firstaid/ health education, wellbeing/fitness initiatives). 7.

Animal Welfare (e.g., animal shelters, walking/grooming pets, support at zoos/ rescues).

8. Community Spaces, Maintenance & DIY (e.g., painting/decorating, playground builds, refurbishments, graffiti removal, repairs). 9. Arts, Culture, Sports & Community Events (e.g., fairs, bakeoffs, sports days/runs, choirs/music, craft/art sessions, cultural activities). 10. Professional Skills & Pro Bono Support (Finance/IT/HR etc.) (e.g., financial literacy and advice, IT/tech help, HR/admin/marketing/legal/data support, coaching/ workshops, board/committee support). 11. Practical Support & Errands for Vulnerable People (e.g., shopping/transport/meal prep/laundry/escorting to appointments). 12. Employee Wellbeing, Team Building & Personal Development (e.g., team bonding/networking, confidencebuilding, freshair/mindfulness/mental health, learning new skills). 13. Community Cohesion, Social & General Wellbeing Activities (e.g. general community meetups, coffee afternoons, cultural sharing events).

29

Blue Sky Thinking: Evaluating the Impacts of Volunteering in Wetland Nature on Creative Thinking and Innovation

For more information contact Lead researcher: Prof. Paul Sowden, paul.sowden@winchester.ac.uk WWT project lead: Dr Jonathan Reeves, jonathan.reeves@wwt.org.uk